Compressing Chemistry Reveals Functional Groups
By: Ruben Sharma, Ross D. King
Potential Business Impact:
Finds hidden patterns in molecules to predict drug effects.
We introduce the first formal large-scale assessment of the utility of traditional chemical functional groups as used in chemical explanations. Our assessment employs a fundamental principle from computational learning theory: a good explanation of data should also compress the data. We introduce an unsupervised learning algorithm based on the Minimum Message Length (MML) principle that searches for substructures that compress around three million biologically relevant molecules. We demonstrate that the discovered substructures contain most human-curated functional groups as well as novel larger patterns with more specific functions. We also run our algorithm on 24 specific bioactivity prediction datasets to discover dataset-specific functional groups. Fingerprints constructed from dataset-specific functional groups are shown to significantly outperform other fingerprint representations, including the MACCS and Morgan fingerprint, when training ridge regression models on bioactivity regression tasks.
Similar Papers
Functional Groups are All you Need for Chemically Interpretable Molecular Property Prediction
Machine Learning (CS)
Helps chemists understand how molecules work.
FGBench: A Dataset and Benchmark for Molecular Property Reasoning at Functional Group-Level in Large Language Models
Machine Learning (CS)
Helps computers understand how molecules work.
Maximizing Efficiency of Dataset Compression for Machine Learning Potentials With Information Theory
Machine Learning (CS)
Finds the most important data to make AI smarter.